20040707Peseau Direct.pdf

Conley E. Ward (ISB No. 1683)
GIVENS PURSLEY LLP
601 W. Bannock Street
O. Box 2720
Boise, ID 83701-2720
Telephone No. [phone]
Fax No. [phone]
cew~givenspursley .com
ECEIVEO.
r... f'ILr.L
""ftn~II H4 Hl~jJUt~ "
j f T l' J t! , ,.tIP ",F: t/ ,Vi Lie
ILl I iL::J LOt""if\-ilSSfON
7/7/t!4/
I? Ese j II/J c-f)
Attorneys for Potlatch Corporation.
S:\CLIENTS\1314\54\Peseau Direct Testimony.DOC
BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION
IN THE MATTER OF THE APPLICATION
OF A VISTA CORPORATION FOR THE
AUTHORITY TO INCREASE ITS RATES
AND CHARGES FOR ELECTRIC AND
NATURAL GAS SERVICE TO ELECTRIC
AND NATURAL GAS CUSTOMERS IN
THE STATE OF IDAHO.
Case Nos. A VU-04-
A VU -04-
DIRECT TESTIMONY OF DENNIS E. PESEAU
ON BEHALF OF POTLATCH CORPORATION
June 21, 2004
ORIG\NAL
PLEASE STATE YOUR NAME AND BUSINESS ADDRESS.
My name is Dennis E. Peseau. My business address is Suite 250, 1500 Liberty Street
, Salem, Oregon 97302.
BY WHOM AND IN WHAT CAPACITY ARE YOU EMPLOYED?
I am the President of Utility Resources, Inc. ("URI"
).
URI has consulted on a number of
economic, financial and engineering matters for various private and public entities for
more than twenty years.
PLEASE DESCRIBE YOUR EDUCATIONAL BACKGROUND AND WORK
EXPERIENCE.
My resume is attached as Exhibit No. 201.
HAVE YOU PREVIOUSLY TESTIFIED BEFORE THE IDAHO PUBLIC UTILITIES
COMMISSION?
Yes, on many occasions.
FOR WHOM ARE YOU APPEARING IN THIS CASE?
I am appearing on behalf of Potlatch Corporation ("Potlatch"
WHAT IS THE PURPOSE OF YOUR DIRECT TESTIMONY?
I have been asked to review A vista s applications in this case and make appropriate
recommendations to the Commission.
PLEASE PROVIDE A SUMMARY OF YOUR TESTIMONY.
My testimony deals with four major issues, all concerning the application for an increase
in electric rates. After reviewing the evidence, I conclude that:
DIRECT TESTIMONY OF DENNIS E. PESEAU - 2
IPUC Case Nos. AVU-O4-1 and AVU-O4-
The Coyote Springs 2 generating plant should be excluded from rate base on
several grounds, not the least of which is that the plant is not "used and useful" in
providing service to A vista s ratepayers.
A vista should not be allowed to recover the cost of natural gas hedges or swaps
put on in April and May of 2001 because they were imprudent and intended to benefit
Avista s unregulated activities at the ratepayers ' expense.
A vista s use of a 2002 test year, adjusted for allegedly known and measurable
changes, produces a mismatch of expenses and rate base, on the one hand, and revenues
on the other. I offer 3 alternative methods of correcting this mismatch.
Avista s inclusion of Potlatch's Lewiston Facility in Schedule 25 for rate design
purposes is unreasonable on its face, and A vista s cost of service study overstates the
annual cost of serving Potlatch by approximately $1.4 million per year.
In addition, John Thornton will present Potlatch's cost of capital testimony and its
recommendation for a return on equity for Avista. However, in the recently completed
Idaho Power rate case, I offered a critique of Dr. Avera s testimony that showed that
updated data and a consistent application of his methodology demonstrate that his cost of
equity is overstated, even if one accepts his assumptions. I fear that if I were to not
perform a similar analysis in this case, the Commission would draw the unwarranted
inference that my critique is no longer valid. To forestall this inference, I have prepared
and attached an Appendix to this testimony that once again shows that simple updates to
Dr. Avera s data, and the use of internally consistent data employed within his return on
equity methods, dramatically lower his return on equity estimate below the 10.4% to
11.9% equity cost range (after the addition of flotation costs) he estimates for benchmark
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. A VU-O4-1 and A VU-O4-
electric utilities in the western U. S., and below the 11.5% equity return he endorses for
A vista.
Coyote Springs 2
WOULD YOU PLEASE EXPLAIN THE ISSUES CONCERNING THE COYOTE
SPRINGS 2 GENERATING PLANT?
Before I do so, a short preface is in order. The two topics I next discuss in this testimony
raise very disturbing issues about the relationship between A vista s regulated and
unregulated arms. In order to understand the significance of these issues, the
Commission needs to have a clear understanding of Avista s peculiar corporate structure.
Consequently, I have reproduced below Scott Morris' Avista organizational chart from
his Exhibit No., page 5 of 5:
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-O4-1 and AVU-O4-
A vista Corporation Company Overview
Avista
Advantage
A vista Energy
Avista Power
- denotes a business entity
- denotes an operating division or line of business
Exhibit No.
S. Morris
Avista Corporation
PLEASE DESCRIBE THE ENTITIES AND OPERATING DIVISIONS ON THE
CHART.
A vista s unregulated enterprises appear on the right hand side of the chart. A vista
Capital is a holding company for these enterprises. A vista Advantage provides
information services and related business services. Neither it nor the operating division
labeled "Other" figure in my testimony. The two entities engaged in "Energy Marketing
and Resource Management " on the other hand, playa prominent role in the following
discussion.
Avista Power is Avista Corporation s ill-fated entry into the merchant power
business. It was originally designed to build or acquire generating plants and other
DIRECT TESTIMONY OF DENNIS E. PESEAU - 5
IPUC Case Nos. AVU-O4-1 and AVU-O4-
resources to serve the unregulated wholesale electricity markets. According to A vista
testimony it is now inactive, but it was the original owner of the Coyote Springs 2
generating plant and it still owns 49% of the Rathdrum merchant plant.
Avista Energy is Avista Corporation s energy trading arm. Its primary purpose is
to trade in both the electricity and natural gas markets. In addition, it brokers deals for
Avista Utilities, although the Washington Utilities and Transportation Commission
recently ordered the termination of this relationship with respect to natural gas purchases.
At the peak of its activity it generated revenues far in excess of A vista Corporation
regulated public utility sales.
YOU EARLIER DESCRIBED A VISTA CORPORATION'S ORGANIZATIONAL
CHART AS "PECULIAR.WHAT DID YOU MEAN?
The right hand side of the chart is not at all unusual for a utility. Most utilities place
unregulated activities in separate entities. The left hand side is quite the opposite. All of
the utilities I am familiar with organize the utility function as a separate business entity,
which makes its own purchases and business deals separate and apart from the
unregulated enterprises. But in A vista s case, there is no separate utility entity, only an
operating division. In effect
, "
A vista Utilities" is simply a name for the residual holder
of A vista Corporation assets that are not claimed by one of the unregulated entities.
WHAT DIFFERENCE DOES A VISTA'S ORGANIZATION MAKE?
It blurs the distinction between regulated and unregulated activities. In the last Avista
rate case, I complained, apparently not strenuously enough, that A vista s corporate
structure, and its practice of not contemporaneously marking trades to its regulated or
non-regulated arm, left it with the latitude to subsequently allocate trades based on their
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. A VU-O4-1 and A VU-O4-
profitability. I characterized this situation as analogous to a stockbroker who makes
investments and then, months or even years later, decides whether the purchases were for
his own or his customer s account.
IS THIS STILL A PROBLEM?
In fact, the present case is far worse. In the case of Coyote Springs 2 ("CS2"), the
unregulated entity (Avista Power) purchased a plant that subsequently proved to be a
disaster. What is the Company s after the fact position? "We (A vista Corporation)
ordered that transaction by our unregulated subsidiary (A vista Power) for the 'benefit' of
our regulated customers." This is analogous to a broker buying a stock for his own
account, and then two years later, when the trade is hopelessly under water, declaring that
the trade was really for the customer s account.
HOW DID CS2 GET STARTED?
The CS2 fiasco began, like many other recent energy debacles in the West, with Enron
playing a prominent role. CS2 was originally a Portland General Electric ("PGE"
project to be built as a companion to PGE's Coyote Springs 1 generating station located
near Boardman, Oregon. PGE was a regulated Enron subsidiary during the entirety of the
CS2 saga.
DID ENRON PLAY ANY ROLE IN THE DEVELOPMENT OF CS2, OTHER THAN
BEING PGE'S PARENT CORPORATION?
Yes. On May 4, 1999 Enron ordered the turbine for CS2 from GE at a contract price of
$35 889 000.
HOW DID A VISTA BECOME INVOLVED WITH CS2?
DIRECT TESTIMONY OF DENNIS E. PESEAU - 7
IPUC Case Nos. A VU-O4-1 and A VU-O4-
In mid-1999, Enron and PGE decided to sell CS2. On October 4, 1999, Avista Power
entered into an "evaluation agreement" with PGE that allowed it to begin its due
diligence investigation of the plant. I assume that other potential buyers were also
investigating the purchase at about the same time.
HOW WAS THE PROPOSED SALE STRUCTURED?
By the time it was completed, the deal was classic Enron in its quirkiness. On October 1
1999, three days before A vista Power signed its evaluation agreement, Enron
incorporated Coyote Springs 2, LLC ("LLC") as a wholly owned subsidiary. On
December 22, 1999, Enron and PGE agreed to transfer CS2 to LLC, contingent upon a
subsequent sale to an unidentified third party. The December 22nd agreement also
divided up the proceeds of the potential sale as follows-both PGE and Enron would first
recover their "cost basis" in CS2 and the turbine, plus their out of pocket and allocated
costs of development. Thereafter, the first $10.47 million of profit was allocated to PGE
the next $12 million to Enron, and any additional amounts were to be split.
DID THIS PGE AND ENRON DEAL CONTEMPLATE A SALE TO A VISTA
POWER?
Not originally. Apparently it was structured for a sale to an unidentified third party who
ultimately backed out. Then A vista Power re-entered the picture. On March 4, 2000
A vista Power signed a Letter of Intent ("LOI") with Enron to buy both CS2 and the
turbine. The LOI set the purchase price at $19.5 million for CS2, and $40 million for the
turbine. PGE's and Enron s collective cost basis and development costs for CS2 were
identified as $ 8 450 000, with the remaining $11 050 000 labeled as a "premium.
WHAT DID A VISTA POWER INTEND TO DO WITH THE CS2 PLANT?
DIRECT TESTIMONY OF DENNIS E. PESEAU - 8
IPUC Case Nos. AVU-O4-1 and AVU-O4-
As in the case of the Rathdrum plant, Avista Power presumably intended to operate CS2
as a merchant plant selling into Western wholesale electricity markets. I base this
presumption in part on the plant's location , which is ill suited to serve A vista Utilities
load centers that are located far to the east of CS2.
DID THE PURCHASE CLOSE AS PLANNED?
No. On June 20, 2000, the LOI was amended to reallocate the purchase price as $16.
million for CS2 and $43 million for the turbine. I cannot find an explanation for this
change in any of the discovery documents we received, although I surmise it may have
been the result of a reduction in the previous estimate of development costs.
An even stranger development took place approximately three weeks later, on
July 7, 2000, when Enron assigned its rights to the GE turbine to Avista Power. On the
same day, Enron created another subsidiary, LJM2-Coyote ("LJM2"). For a price of
540 000, LJM2 provided Avista Power with a two week "put option" on the turbine.
In other words, from July 7th through July 21st, Avista Power could require LJM2 to
repurchase the turbine for the sum of $39 960 000. This put option was never exercised
because, on July 21 , 2000, Enron assigned its interest in LLC to Avista Power, thus
giving A vista Power ownership of CS2 as well as the turbine.
WHY IS THE LJM2 TRANSACTION STRANGE?
I can think of no legitimate business reason for A vista Power to enter into the put option
agreement. In the first place, turbines were in short supply at the time, and A vista would
have had little difficulty re-selling the turbine if the CS2 deal somehow collapsed.
Moreover, it is difficult to understand why, if Avista Power feared the exposure of
holding the turbine before it secured the CS2 rights, it didn t simply insist on a
DIRECT TESTIMONY OF DENNIS E. PESEAU - 9
IPUC Case Nos. AVU-O4-1 and AVU-O4-
simultaneous transfer of the two components. Instead it allowed Enron to impose a two-
week gap on the signing of the two agreements and, in effect, sell it $3.5 million of
insurance to cover the minimal exposure that gap created. Finally, why would any
reasonable businessperson pay $3.5 million for a two week "insurance policy" issued by
an empty corporate shell, with no assets and an operating history of less than a day, even
if Enron guaranteed the put? This simply doesn t pass even a minimal smell test
particularly when the counter party is named Enron.
WHEN ALL WAS SAID AND DONE, WHAT DID A VIST A PAY FOR CS2 AND
THE TURBINE?
The total purchase price, including the option, was approximately $59.5 million, for a
plant that, by my calculations, appeared to have an all-in cost of approximately $42
million.
WHAT WAS THE BOOK VALUE OF THE TRANSFERRED ASSETS?
The book value of the turbine would have been the same as its purchase price
$35 889 000. The Allocation Agreement dated July 21 , 2000 listed CS2's book value as
755 409, with an additional $2 287 591 allocated to project development expenses.
Consequently, the book value would have been $39 644 409 without the development
expenses, and $41 932 000 if development expenses were capitalized and added to book
value.
WAS THAT THE END OF A VISTA POWER'S INVOLVEMENT WITH ENRON?
Not quite. In April of 2002, CS2's prime contractor, another Enron affiliate, filed for
bankruptcy and CS2 lost the benefit of its fixed price construction contract, while at the
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. A VU-04-1 and A VU-04-
same time incurring the cost of replacing the prime contractor and settling with
subcontractors.
WAS THAT THE ONLY PROBLEM THAT OCCURRED DURING THE
CONSTRUCTION AND OPERATION OF CS2?
No. It is fair to say that CS2 has been, and continues to be, an economic and operational
nightmare. In May of 2002, approximately a month before the scheduled completion of
the plant, a fire destroyed the transformer purchased from a Turkish supplier. This not
only prevented the completion of the plant, it also resulted in an environmental incident
when water used to douse the fire overran the splash pond built to contain the
transformer s contents in the event of an accident. Clean-up costs as of December 31
2003 were approximately $1.7 million, half of which are A vista s responsibility.
A replacement transformer arrived at the site in December, 2002, but an
inspection revealed it could not be installed because of shipping damage. Repairs to this
transformer delayed CS2's commercial operation date for more than a year, to July, 2003.
Thereafter, the plant was in service for approximately six months. It then experienced
another round of transformer problems that shut it down again. The projected date for a
return to service is now August of2004.
YOU JUST DESCRIBED CS2 AS AN ECONOMIC NIGHTMARE. ARE YOU
REFERRING TO SOMETHING BEYOND ITS CONSTRUCTION PROBLEMS?
Yes. The construction problems have caused the estimated cost of A vista s half of the
plant to swell from approximately $94 million to $109 million. In addition, the natural
gas swaps I will discuss in detail later in my testimony produced losses in excess of $62
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
million. The bottom line is that A vista overpaid for the plant in the original purchase
and every turn of the cards since then has only added to the misery.
so WHO PAYS FOR ALL THIS?'
Under Avista s proposal to rate base the entirety of the plant's cost , Avista ratepayers
will pay for all of these problems. If A vista s proposal is accepted, the only entities that
walk away from this train wreck unscathed are the plant's original owner , Avista Power
and its parent, A vista Corporation.
HOW DOES A VISTA POWER ESCAPE ANY RESPONSIBILITY FOR CS2'
PROBLEMS?
In December of 2000, A vista Corporation announced it would acquire CS2 from A vista
Power. But it did not in fact follow through on this announcement. Instead, it vacillated.
Internal A vista memos indicate that A vista Power was trying to sell the entire plant to
third parties in the summer and fall of 2001. But A vista Power received only one full
price offer from Mirant, and that prospective deal fell apart when Mirant ran into cash
flow problems. Ultimately, Avista Power ended up selling 50 percent of the plant to
Mirant, and 50 percent to Avista Corporation.
WHEN DID THESE SALES OCCUR?
A vista Power assigned a 50 percent interest in LLC to Mirant on December 12, 2001.
But it did not transfer the other 50 percent of the plant to Avista Corporation until
January 1 , 2003 , after the close of the test year in this case.
GIVEN THIS HISTORY, WHAT IS THE APPROPRIATE RATEMAKING
TREATMENT FOR CS2?
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
I have two recommendations concerning CS2. The first is that the cost of the plant
should not be included in rate base in this case. CS2 is demonstrably not used and useful
and its track record does not inspire confidence it will be used and useful in the near
future. A vista has had three tries at completing the plant and getting it running on a
reliable basis. It has struck out all three times. Given this history, the plant's costs
should not be eligible for recovery in regulated rates until it has a demonstrated track
record of usefulness and reliability.
Furthermore, if and when the plant does become eligible for inclusion in rate
base, the rate based costs should be limited to the plant's fair market value, as described
below, as of the transfer date of January 1 , 2003.
WHY ARE YOU PROPOSING TO REDUCE THE PLANT'S COST IN THIS
MANNER?
I am simply applying standard ratemaking precepts to the purchase. A vista Power is an
unregulated Avista Corporation subsidiary, and transactions between it and Avista
Corporation are clearly not at arms length. I am not an attorney, but I have spent enough
years in the regulatory field to state that, in jurisdictions I am familiar with, when a utility
purchases goods or services from an unregulated affiliate, the burden is on the utility to
prove that the purchase price did not exceed fair market value. In the present case
because of all the construction disasters, it is quite clear that transferring CS2 to A vista
Corporation at cost creates a purchase price that is well in excess of fair market value.
These excess costs should be disallowed. It is patently unjust to ask the
ratepayers to relieve A vista Power of the unfortunate consequences of its half ownership
of CS2.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 13
IPUC Case Nos. AVU-04-1 and AVU-04-
DOES THE FACT THAT A VISTA CORPORATION PREVIOUSLY ANNOUNCED
AN INTENTION TO ACQUIRE THE PLANT MAKE ANY DIFFERENCE IN THIS
CASE?
No. Avista s announced intentions came after Avista Power had already overpaid for the
assets it purchased from PGE and Enron, so an adjustment to fair market value would
have been in order even then. In addition, even though the boards of directors of the
involved companies authorized their executives to proceed with the transaction, the
companies never acted on those resolutions. A vista s discovery responses contain no
contract, memorandum of understanding, or any other document that would evidence an
intention to proceed with the sale. Under those circumstances, A vista Power was under
no legal obligation to sell to A vista Corporation, and it in fact tried to sell the plant to
third parties months after the announcement. Eventually it did sell half to Mirant.
Avista unilaterally chose to purchase CS2 through its unregulated subsidiary,
thereby avoiding any regulatory constraints on its use or disposition of the assets. Let
suppose that A vista Power had succeeded in the summer of 200 1 in selling the plant at a
profit. Would A vista Power have volunteered to share the proceeds with the ratepayers
just because at one time it intended to sell the plant to Avista Corporation? This is the
same A vista that resisted sharing the Centralia sale proceeds with ratepayers. A vista
would have argued that the deal was never consummated, and that ratepayers never
acquired an equitable interest in the plant through the payment of depreciation.
HOW DO YOU PROPOSE TO DETERMINE THE FAIR MARKET VALUE OF CS2?
The Commission could conduct further proceedings for the express purpose of making
such a determination, but there is a much easier method readily available. Just two years
DIRECT TESTIMONY OF DENNIS E. PESEAU - 14
IPUC Case Nos. AVU-04-1 and AVU-04-
ago, the Commission conducted an extensive investigation to determine the cost of a 270
megawatt combined cycle natural gas plant to use as a surrogate avoided resource
SAR") for the purpose of calculating avoided cost rates. On August 2, 2002, one
month after CS2' s original scheduled completion date, and five months before the
transfer of CS2 to A vista Corporation, A vista filed rebuttal testimony identifying the
most recent construction cost estimates for the SAR as $604/kilowatt. I see no reason
why A vista should not be held to its own contemporaneous estimate of the cost of
constructing a plant nearly identical to CS2. This figure, after all, represents the
maximum value A vista Corporation was willing to pay for the purchase of resources
from unrelated third parties just before it acquired CS2 from Avista Power. Using the
$604 figure produces a fair market value for CS2 of $84 560 000 for A vista s share of
CS2. The Commission should not allow costs above this amount in rate base at any time.
The Natural Gas Hedges
WHAT IS THE ISSUE WITH RESPECT TO THE "DEAL A" AND "DEAL B"
HEDGE TRANSACTIONS IN THE COMMISSION'S ORDER ON A VISTA'S 2003
PCA FILING?
To its credit, the Commission recognized the peculiar nature of both Deal A and Deal B
in the 2003 PCA proceeding and deferred a decision on the costs of these deals into the
present general rate case. As I explain below, the high costs associated with each deal are
the result of imprudent decisions and self-dealing between A vista Corporation and A vista
Energy. Avista s actions have resulted in excess natural gas costs of more than $62
million on a system-wide basis.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 15
IPUC Case Nos. AVU-04-1 and AVU-04-
HAVE MOST OF THE INFORMATION, DATA, AND FACTS NECESSARY TO
UND ERS T AND THE NATURE OF DEAL A AND DEAL B BEEN TREA TED AS
CONFIDENTIAL BY AVISTA?
Yes. This is unfortunate, as most of the confidential trading data necessary to understand
Deal A and Deal B are public and available on the FERC website as part of the FERC'
show-cause proceeding that culminated in its March 2003 P A02-02 report Final Report
on Price Manipulation in Western Markets There is, therefore, no valid reason to
continue to treat historical trading data as confidential.
WHAT IS THE DIFFERENCE BETWEEN THE NATURAL GAS TRANSACTIONS
OF DEAL A AND DEAL B AND NORMAL NATURAL GAS TRANSACTIONS?
There are at least three distinct aspects of the Deal A and Deal B transactions that are
peculiar. The first distinction is that the Deal A and Deal B trades were financial as
opposed to physical transactions.
WHAT IS THE DISTINCTION BETWEEN NATURAL GAS FINANCIAL AND
PHYSICAL TRANSACTIONS?
A physical transaction is the more normal and common purchase of an actual, physical
quantity of natural gas at specified pricing, terms and conditions. In physical gas
transactions, there are no winners or losers. The buyer receives a specific quantity of gas
at agreed upon pricing terms. The seller receives a payment for providing the physical
gas to the buyer.
A financial natural gas transaction involves no actual exchange of physical gas.
Instead, a financial deal is agreed upon by buyer and seller in which the buyer bets that
future gas prices will increase, while the seller bets that future gas prices will decrease.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 16
IPUC Case Nos. AVU-04-1 and AVU-04-
Depending upon the future monthly movement of gas prices, the loser, or the
counterparty on the wrong side of the bet writes a monthly check or "settles" with the
other party. The FERC report just referenced defines financial gas swaps similar to Deal
A and Deal Bas:
In a swap, two counterparties execute a trade in which the buyer pays a
fixed, known price for some notional quantity of gas and the seller pays a
price that will vary with the market price (generally based on some agreed
upon price index), which will only be known later. Thus, the buyer in a
swap transaction is going long making a bet that the market price will
rise - and the seller is betting that prices will fall.
(Page II-51)
On the four days April 10, April 11 , May 2 and May 10, 2001 , A vista Energy
entered into the financial swaps, Deal A and Deal B, on behalf of A vista Utilities that
were of unprecedented length and lost over $62 million for ratepayers. At no time during
the terms of these two deals were these financial trades "in the money," or profitable for
A vista Utilities. The deals were extraordinarily profitable for the three seller
counterparties.
WHO WERE THE COUNTERPARTIES TO THESE TRANSACTIONS?
BP and Mirant were the counterparties on Deal A. Incredible as it may seem, A vista
Energy was the counterparty for Deal B.
WHY WOULD THE SAME ORGANIZATION SIMUL ANEOUSL Y TAKE
OPPOSITE SIDES OF THE BET ON THE DEAL B SWAP? ISN'T THIS A "ZERO
SUM GAME?'
The fact that the PCA protected A vista Corporation is the only thing that made this
attractive transaction for A vista Corporation. The PCA insulated the shareholders of the
DIRECT TESTIMONY OF DENNIS E. PESEAU - 17
IPUC Case Nos. A VU-04-1 and A VU-04-
parent company by passing through to ratepayers the excess of the locked in hedged
natural gas prices over and above the actual market prices that existed at the time.
MIGHT THIS BE SIMPLY A CASE OF BAD LUCK FOR A VISTA'S CUSTOMERS?
No. The only manner in which a financial swap can be consummated is with a willing
buyer and a willing seller. The reason for entering a swap on either side is because one
information on market pricing makes the risk of this bet worthwhile. Again, the only
possible reason for A vista Utilities to buy the long-term financial swap that it did was
because it was predicting gas prices would continue to increase. If future gas prices at
the time the swap was entered were expected either to remain at the then high levels, or to
decrease then entering the fixed price swap could only harm the buyer. On the other side
the seller A vista Energy apparently had information suggesting that future gas prices
were not going to rise above the agreed upon price per decatherm over the subsequent
months, or it would have been foolish to sell the swap. Unless Avista Energy based its
action on information that prices would either remain at their high levels or fall, it would
have been acting directly against the best interests of its shareholders. If natural gas
prices truly were expected to increase over the subsequent 17 months, the best action for
both A vista Utilities and A vista Energy would have been for A vista Utilities to buy the
fixed-price swap from a less informed counterparty.
IS THERE ANYTHING ELSE UNUSUAL ABOUT A VISTA CORPORATION'
DECISION TO MAKE THE SWAP?
Yes. At the time, A vista Energy brokered all of the natural gas and electric trades made
for the benefit of A vista Utilities. A vista s justification for this practice was that A vista
Energy s continuous market participation provides it with market insights and knowledge
DIRECT TESTIMONY OF DENNIS E. PESEAU - 18
IPUC Case Nos. AVU-04-1 and AVU-04-
that the utility division doesn t have. Avista Energy s role as a broker for the utility
division placed it in a fiduciary position that required it to disclose the fact that it
considered Deal B (and by implication, Deal A) to be a bad deal for Avista Utilities.
A vista Energy did disclose that fact and the additional fact that it was taking the other
side of the swap, it was obviously imprudent for A vista Utilities to proceed with swaps
that the party with superior knowledge regarded as foolish. If A vista Energy did not
disclose its role, then it violated its fiduciary responsibilities, and that alone would be
grounds for disallowing the cost of both deals in rates.
WHAT WAS THE RESULT OF THE DEAL B SWAP WITH A VISTA ENERGY?
The result was that A vista Utilities immediately began monthly transfers of what turned
out to be millions of dollars to A vista Energy.
HOW COULD THERE BE AN IMMEDIATE TRANSFER OF CASH? I THOUGHT
THE SWAP WAS FOR GAS TO BE DELIVERED IN THE FUTURE.
The immediate impact occurred because of the way financial trades such as this are
settled. As I stated earlier, swaps like this are literally bets on the direction of prices.
Consequently, they settle monthly based on the futures price of gas for the time period
covered. In any month in which the futures price is less than the fixed price, the buyer
(Avista Utilities) loses his bet and must cut a check to the seller (Avista Energy) for the
difference.
WHAT IS THE ULTIMATE SIGNIFICANCE OF THE WAY THESE TRADES ARE
SETTLED?
1 A vista converted Deal B to a physical purchase at an equivalent fIXed price on June 20, 2002.
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
It explains why the Commission really has no choice but to disallow Deal B. Any other
decision would provide Idaho utilities that have a PCA or PGA with a blueprint on how
to raid ratepayers' pockets for the benefit of shareholders.
HOW DOES A VISTA UTILITIES ATTEMPT TO JUSTIFY ITS DECISION TO
ENTER INTO "BUYS" IN BOTH DEAL A AND DEAL B?
Avista witness Mr. Lafferty discusses these two deals (actually four transactions) in
pages 29-56 of his testimony. The attempted justification, while sometimes repetitive, is
outlined as follows: Deal A and Deal B were made because:
1. A vista was in an electric resource deficit or a "short-position" during the hedge
periods. (pp. 31-, 37-42-47)
2. The high hedge prices of Deal A and Deal B still compared favorably to forward
market prices of electric purchases at the time. (pp.32-36)
3. Electric market prices in January-May 2001 were high, and federal opposition to
price caps suggested no relief in market prices. (pp. 40-, 41-42)
4. The 36 month and 17 month duration of Deal A and Deal B were not unusual terms
for company hedges of this sort. (pp. 48-52)
5. The company did not expect that forward natural gas prices would decline as they
did. (pp. 52-53)
6. The terms of Deal A and Deal B were consistent with market conditions on April 10
and May 10. (pp.53-54)
WOULD THE DEFICIT ELECTRIC RESOURCE POSITION IDENTIFIED BY THE
COMPANY JUSTIFY BUYING FINANCIAL HEDGES LIKE DEAL A AND DEAL
DIRECT TESTIMONY OF DENNIS E. PESEAU - 20
IPUC Case Nos. A VU-04-1 and A VU-04-
No. I first want to make clear that Potlatch does not want in any way to discourage
appropriate resource acquisitions to maintain the reliability of service to customers.
However, I am quite surprised that the company testimony in this regard suggests that
somehow Deal A and Deal B in any way assisted in covering a resource-short position.
WHY DO YOU INDICATE THAT DEAL A AND DEAL B DID NOT ASSIST
VISTA IN COVERING ANY RESOURCE DEFICIT?
Financial fixed-for-floating swaps such as Deal A and Deal B never provide for any
physical quantities of natural gas. Again, Deal A and Deal B are strictly the taking of
price positions" between two parties, a buyer and seller. For example, if I thought that
natural gas prices were going to increase in the near-term, and I could locate a party
thinking the opposite, I could buy a natural gas financial swap and reap gains or suffer
losses according to my accuracy, and never be involved with actual physical quantities of
gas.
If I need natural gas to close an electric resource deficit, I would need to enter into
distinct physical gas contracts as a buyer. Deal A and Deal B did not entitle Avista to
even a molecule of methane.
IF A VISTA NEEDED ADDITIONAL NATURAL GAS SUPPLY TO COVER THE
PERCEIVED DEFICIT, HOW DID IT ACQUIRE SUCH SUPPLIES?
The company on March 13 and March 22, 2001 , entered into 36 month and 17 month
physical trades for 27 658 and 20 000 decatherms per day at market index-based prices.
These two gas contracts alone filled the need to cover the resource deficits discussed by
the Company. Deal A and Deal B merely expressed the perceived direction that natural
gas prices would take over the ensuing 36 and 17 month periods between the betting
DIRECT TESTIMONY OF DENNIS E. PESEAU - 21
IPUC Case Nos. AVU-04-1 and AVU-04-
2 '
parties. The Commission should reject any notion that these financial swaps can be
peddled to customers on the basis of enhancing system reliability.
WHAT DO YOU MAKE OF MR. LAFFERTY'S DISCUSSION ON PAGES 32-36 OF
HIS TESTIMONY THAT SUGGESTS THE DEALS WERE PRUDENT BASED ON
THE THEN FORWARD MARKET PRICES?
The analysis at pages 32-36 of Mr. Lafferty s testimony attempts to demonstrate that the
variable cost of power produced by A vista s generators would have been below the
predicted future market power prices at the gas prices in Deal A and Deal B. That is
A vista was predicting that at the Deal A and Deal B fixed swap prices, buying gas for
internal generation would be cheaper than buying on the electric markets. This assumes
of course, that the existing forward power prices at mid-Columbia represented a good
predictor of actual prices in the future.
While this analysis is mathematically correct, it hardly demonstrates that the Deal
A and Deal B trades were prudent.
PLEASE EXPLAIN.
The analysis presented is the starting point for an "arbitrage" trade. An arbitrage is the
simultaneous buying and selling of fungible commodities in different markets in order to
make an immediate riskless profit. For clarification of the proper use of Mr. Lafferty
analysis I refer to the Coyote Springs 2 table at the bottom of page 32 of his testimony.
The first row indicates that the Deal B gas fixed price is $6.56 per decatherm and, at the
CS2 plants' heat rate , Deal B gas could produce electricity at a variable cost of
$46.06/MWh. The forward electric prices, according to Avista, showed power prices at
the time of$126.75 and $105.38/MWH.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 22
IPUC Case Nos. A VU-04-1 and A VU-04-
A power trader facing these circumstances would, if the market held
simultaneous lock in a buy at the $6.56 gas price and a sale at the $126.75 and
$105.3 8/MWh electric prices to insure a riskless profit equal to the difference between
these two energy sale prices and the $46.06/MWH the electricity would cost to produce.
This would be a rational use of Mr. Lafferty's analysis.
DOES THE ANALYSIS PRESENTED BY MR. LAFFERTY DEMONSTRATE THAT
DEAL A AND DEAL B WERE PRUDENT AT THE TIME FOR THE PURPOSE OF
PROTECTING RATEPAYERS?
No. Unlike the arbitrage case where a certain outcome (the riskless profit) is locked in by
a conscious decision to forego possible upside and avoid all downside, the open hedges
conducted by A vista did the opposite. A vista s hedges in essence locked in the downside
- by fixing gas prices at near record levels for up to 36 months - and precluded the
ratepayers getting any upside if gas prices returned to more normal historic levels.
WOULD A VISTA ENERGY HAVE ENTERED THE SELL SIDE OF THESE
HEDGES IF IT EXPECTED NATURAL GAS PRICES TO CONTINUE UPWARD?
Absolutely not. Doing so would have been a direct contradiction of management's
fiduciary responsibility to shareholders. A vista Energy made a calculated bet that the
very high natural gas market prices could not be sustained. By selling Deal B to the
utility for prices that exceeded $6.00/decatherm it stood to reap all the profit from falling
prices. If prices simply remained at the then high levels, A vista Energy stood to lose
nothing. Only if gas prices increased further from these high levels, did it risk losing
money. The end result is that Avista Energy made an obvious bet and reaped more than
$18 million in benefits from its parent utility.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 23
IPUC Case Nos. AVU-04-1 and AVU-04-
PLEASE ADDRESS MR. LAFFERTY'S DISCUSSION ON PAGES 40-
REGARDING THE PRUDENCE OF THESE TRANSACTIONS.
Beginning on line 17 of his page 40, Mr. Lafferty suggests that a prudent person would
have viewed the high winter prices of2000-2001 , and the federal government's position
against the implementation of price caps, as reasons to "go long" with the natural gas
hedges. I have just two short comments on this point.
First, the prudent man at A vista who was buying the fixed-price hedge on behalf
of the utility was the same man who was selling it on behalf of Avista Energy. Taking
simultaneous and opposite positions on the same transaction cannot each be deemed
prudent. The same market observation of high prices and price caps could not have led a
single individual or committee to opposite conclusions regarding the future near-term
trend in gas prices.
Second, other utilities and market participants in the western U.S. observed the
same market phenomena discussed by Mr. Lafferty and did not take long-term price
positions that anticipated further gas price increases.
PLEASE DISCUSS MR. LAFFERTY'S TESTIMONY ON PAGES 48-52 THAT
SUGGESTS THAT THE 36 MONTH AND 17 MONTH HEDGES ARE COMMONLY
MADE BY THE UTILITY.
Mr. Lafferty s discussion here involves only physical resource acquisitions, not financial
hedges. I certainly agree with him that any resource portfolio should have various short
medium, and long-term resources. In this light, I do not challenge or take issue with
Avista s entering into its March 13 and March 22 long-term physical gas purchase
contracts, as I previously noted.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 24
IPUC Case Nos. A VU-04-1 and A VU-04-
The issue here, of course, is that A vista took an unprecedented long-term price
view in the form of financial hedges and, in combination with its subsidiary A vista
Energy, A vista Corporation took both sides of the transaction. Mr. Lafferty is silent on
these points.
HAS A VISTA EVER, TO YOUR KNOWLEDGE, ENTERED INTO FINANCIAL
HEDGES AS LONG AS THE 36 MONTH AND 17 MONTH TERMS OF DEAL A
AND DEAL B?
No. In response to Potlatch's data requests, Avista provided a list of all recent financial
hedges and fixed price contracts. Of the 67 fixed-price transactions provided, the
overwhelming majority of the contracts were for terms of 1-3 months, with a few with
terms of one year. Only the Deal A and Deal B transactions were for such long periods.
I conclude that it is not Avista s normal business practice to enter into long-term price
hedges.
HAVE YOU REVIEWED OTHER DATABASES FOR INFORMATION TO
DETERMINE WHETHER THE 36 AND 17 MONTH TERMS OF DEAL A AND
DEAL B ARE COMMONPLACE IN THE INDUSTRY?
Yes. In conjunction with its investigation of electric and natural gas price manipulation
in western U.S. markets, the FERC compiled massive databases regarding both physical
and financial natural gas trades. As a check on the frequency of long-term financial
hedges, I reviewed the FERC data file for all natural gas financial hedges that were
entered into during May 2001 , the same period as Deal A and Deal B.
According to the data base file, there were 37 472 such transactions during May
2001. The huge preponderance of these financial hedges was for the immediate month or
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
quarter ahead, although some were for quarterly periods ending as late as December
2002. I found no other financial trades that extended as long as the 36 and 17 month
terms contained in Deal A and Deal B.
PLEASE ADDRESS MR. LAFFERTY'S TESTIMONY THAT THE DECLINE IN
NATURAL GAS PRICES WAS UNFORESEEABLE.
Mr. Lafferty s testimony on pages 52-53 states that "the Company" did not expect that
forward natural gas prices would decline, as of course they did (Page 52, lines 3-6). I
cannot from the context of the statement ascertain just what "the Company" is. Certainly,
A vista Energy expected a decline in natural gas prices, or it would not have sold the fixed
prIce swap.
Further, Mr. Lafferty s explanation does not justify the utility buying the swap.
As I explained earlier, buying the fixed-price swap only gave the utility protection from
further increases in gas prices, not from the then existing level of high prices. Mr.
Lafferty explains only that" .
. .
the Company expected the price for natural gas would
remain high for some time into the future..." (page 52, lines 5-6). He does not make the
argument that the Company expected gas prices to continue to increase, which would be
the only legitimate reason for the swaps.
WERE THE TERMS OF DEAL A AND DEAL B CONSISTENT WITH MARKET
CONDITIONS ON APRIL 10 AND MAY 10 2001 , AS MR. LAFFERTY ARGUES?
As I have previously indicated, there were apparently no other natural gas hedge
transactions occurring that were comparable to Deal A and Deal B. The references Mr.
Lafferty makes to forward price curves at that time certainly is no indication of what an
DIRECT TESTIMONY OF DENNIS E. PESEAU - 26
IPUC Case Nos. AVU-04-1 and AVU-04-
arms-length buyer and seller might agree upon for financial hedges of up to 36 months in
length.
WHAT IS YOUR RECOMMENDATION WITH RESPECT TO THE FINANCIAL
LOSSES CLAIMED BY THE UTILITY IN CONJUNCTION WITH DEAL A AND
DEAL B?
The financial losses incurred by the utility in Deal A and Deal B are summarized in my
Exhibit No. 202. As of March 31 , 2004, the cumulative losses to the utility on the hedges
were $62 446 000. These losses represent the difference between what the utility would
have paid for natural gas on the market (absent the hedges) and the high fixed gas price
that it agreed to pay by virtue of the hedges. The market prices for gas are shown for the
Malin receipt point, and are compared to the weighted average price of the hedges
labeled "Average $/dt." For Deal A, the cumulative financial loss was $44 175 600. For
Deal B , the cumulative loss was $18 270 400.
Since Deal B involves self-dealing and a direct transfer of the utility s losses to
shareholder profits, the entire $18.3 million must be disallowed, adjusted of course for
the Idaho jurisdictional share and for the PCA test period. Deal A did not involve self
dealing, but it was certainly imprudent and it is further suspect due to the unprecedented
term of 36 months and the high locked in prices. I believe it should likewise be
disallowed. But if the Commission for some reason rejects this proposal, I propose, in
the alternative, a lesser adjustment based on a more normal hedging strategy.
PLEASE EXPLAIN THE LATTER RECOMMENDATION.
Deal A represents two hedge contracts of 10 000 decatherms each for a period of 36
months. The named counter parties to these Deal A contracts are private entities with no
DIRECT TESTIMONY OF DENNIS E. PESEAU - 27
IPUC Case Nos. A VU-04-1 and A VU-04-
apparent legal connection to A vista. According to the Company s response to Potlatch'
data requests, A vista did not have either of these entities "sleeve " (conduct the trade for
Avista Energy s benefit) the transaction. Thus, there was no apparent enrichment of
Avista s shareholders. But Deal A was nevertheless an imprudent $44.2 million hedge
given its duration and the fact that it was put on contrary to A vista Energy s position.
I base my adjustment on A vista s normal hedge strategies for all its other fixed
price gas purchases. As I stated earlier, Avista normally hedges for gas deliveries in
ensuing seasons and occasionally for periods as long as one year. If Avista had followed
its normal hedging strategy it would have avoided the disastrous 36 month Deal A fixed
price of $6.45/decatherm.
HOW IS THIS INFORMATION USED TO CALCULATE AN ADJUSTMENT FOR
DEAL A?
My review of A vista s confidential information on other hedges reveals that A vista
normal hedges were established approximately six months prior to a season (November-
March or April-October). I therefore used the Malin natural gas contract prices in effect
six months prior to each upcoming season as a base price. For example, May 1 , 2001
prices were used for the November 2001-March 2002 season. These prices are then
subtracted from the Deal A prices. The results are summarized in my Exhibit No. 203.
WHAT DOES EXHIBIT NO. 203 SHOW?
That exhibit indicates that, if A vista had not entered into Deal A and instead hedged in
the same manner that it was hedging other natural gas purchases in the same time frame
gas costs would have been $30 365 240 lower. I alternatively propose that, should the
Commission not disallow the entirety of the Deal A costs, it should disallow $30.4
DIRECT TESTIMONY OF DENNIS E. PESEAU - 28
IPUC Case Nos. AVU-04-1 and AVU-04-
million of Deal A costs, adjusted for both the Idaho jurisdiction as well as the PCA test
period.
The Test Year Mismatch
YOU EARLIER STATED THAT A VISTA'S CASE CONTAINS A MISMATCH OF
REVENUES AND EXPENSES. PLEASE EXPLAIN WHAT YOU MEAN BY THE
WORD "MISMATCH.
Avista calculates its test year revenues in a straightforward manner. Test year revenues
consist of 2002 actual figures
, "
normalized" for weather and other standard Commission
approved adjustments. On the other side of the ledger, however, expenses and rate base
are treated in a much different manner. A vista pro forms increases in selected expense
items, such as pension, insurance, and labor costs, to 2004 levels. A vista also includes in
rate base a number of projects that were placed in service after the test year, as well as
construction work in progress that is scheduled for completion in 2004. These
adjustments produce operating and maintenance increases of approximately $5.4 million
rate base additions of $54 million, and associated depreciation increases of $2.3 million.
The net effect is a mismatch of 2002 revenues against year-end 2004 expenses and rate
base.
IS THIS AN ACCEPTABLE RA TEMAKING PROCEDURE?
No. For unknown reasons, Avista chose a 2002 test year, rather than 2003. Having made
that choice, it should not be allowed to unilaterally alter the test year relationship between
revenues, expenses and rate base. It is a fundamental principle of regulation that a
utility's rate base and expenses should be matched against revenues for the same period.
A vista s pro forma results clearly violate this principle.
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
ARE YOU SUGGESTING PRO FORMA CHANGES TO A TEST YEAR BASE CASE
SHOULD BE REJECTED OUT OF HAND?
No. Adding known and measurable changes to a test year base case is a legitimate
regulatory tool, but it must be used with extreme caution because of the high potential for
abuse. In a rate case, utilities have every incentive to identify changes that increase the
revenue requirement, but no incentive at all to find revenue enhancing changes.
Consequently, it comes as no surprise that all of Avista s proposed known and
measurable changes produce an increase in revenue requirement. These changes should
either be rejected or accompanied by a corresponding adjustment to revenues.
CAN YOU PROVIDE AN EXAMPLE OF THE TYPE OF KNOWN AND
MEASURABLE CHANGE THAT SHOULD BE ACCEPTED?
The classic example is a post-test year change in tax rates. Plugging the new tax rates
into the revenue requirement calculation does not disturb the relationship between test
revenues and expenses and is obviously equitable.
WHAT RULES SHOULD BE APPLIED TO POST-TEST YEAR KNOWN AND
MEASURABLE CHANGES?
Post-test year expense and rate base adjustments should only be accepted when they are
in fact truly known and measurable. In order to qualify, a proposed adjustment must be
virtually certain to occur, and its revenue requirement impact must be precisely and
reliably quantifiable. Furthermore, there must be some limit on the time interval between
the test year and pro forma adjustments.
ARE A VISTA'S PRO FORMA ADJUSTMENTS CONSISTENT WITH THE RULES
YOU HAVE JUST DESCRIBED?
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. A VU-04-1 and A VU-04-
No. In the case of its pro forma expense adjustments, the time lag between the 2002 test
year and adjustments based on 2004 data or projections makes these adjustments
inequitable.
WHY IS THE TIME LAG IMPORTANT?
For most utilities, expenses tend to increase every year, but this is offset in whole or in
part by efficiency improvements and load growth. If this were not so, utilities would
automatically file rate cases every year. Avista s own rate case history nicely illustrates
this point. Its last rate case occurred in 1998, and the one before that was several years
earlier.
Avista s pro forma expense adjustments for items like increased labor, insurance
and similar costs are simply 2004 budget estimates. It is absolutely inappropriate to
match these expenses against 2002 revenues because normal load growth will recoup
some or all of these costs. The Commission should either reject the 2004 adjustments or
impute revenue increases to the 2002 test year to correct this mismatch.
ARE A VISTA'S PRO FORMA ADDITIONS TO RATE BASE SUBJECT TO THE
SAME OBJECTIONS?
Only in part. Additions to A vista s generating capacity were added to the power supply
model, and this presumably adds revenues or decreases expenses as a result of the pro
forma plant additions. I have not attempted to confirm that this modeling change was
properly implemented, but I assume Staff will do so. If the implementation was correctly
done, I have no objection to these pro forma adjustments as such, although I have
proposed the removal of Coyote Springs 2 on other grounds, as discussed above.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 31
IPUC Case Nos. AVU-04-1 and AVU-04-
But there is no similar revenue adjustment for the $26 300 000 in 2003 and 2004
transmission projects A vista pro forms into the rate base, even though these additions will
also produce either additional revenues or operational savings. Like other businesses
utilities generally do not make additional investments or increase their expenses unless
they can generate additional revenues and profits, either by serving additional customers
or by cutting costs or increasing margins. There is no reason to assume this is not the
case here. The projected expenditures Avista has identified must be presumed to
generate additional revenues or other benefits that would offset their costs, in whole or in
part. A vista has made no attempt to identify these offsetting benefits.
As the Commission pointed out in its recent order in the Idaho Power rate case:
Generally speaking, the Commission expects all utilities to attempt to identify
expense saving and revenue producing effects when proposing rate base
adjustments for major plant additions. It is unfair to ratepayers to assume that the
investment in these plants will not increase Company revenues or decrease
Company expenses in the future. Further, it is unreasonable to expect the
Commission to allow full recovery of plant investment as if the plant had been in
operation the full year without a corresponding adjustment to revenues and
expenses.
Order No. 29505 , p. 7.
HOW SHOULD THIS MISMATCH BE CORRECTED?
There are basically three alternative remedies available to correct this rate base mismatch.
The first would be to reverse the pro forma entries and properly match test year averages
on both sides of the ledger. The second alternative is to update revenues to the 2004 level
in the same manner as rate base and expenses. Finally, the third alternative is to employ
the rate base adjustments the Commission adopted in the Idaho Power rate case.
DO YOU HAVE A PREFERENCE BETWEEN THESE THREE AL TERN A TIVES?
DIRECT TESTIMONY OF DENNIS E. PESEAU - 32
IPUC Case Nos. A VU-04-1 and A VU-04-
As I have stated in other cases, I think annualizing revenues to 2004 year-end levels is the
preferable course for two reasons. First, it is much simpler to annualize revenues than to
back out pro forma adjustments from numerous expense and rate base categories.
Moreover, adjusting revenues produces a more forward-looking result than reversing the
expense and rate base annualizations.
I recognize, however, that the Commission adopted a third course of action to
correct similar mismatches in the recent Idaho Power rate case. In that case, the
Commission adopted a proxy for increased revenues and reduced expenses. While the
Commission stated that it did not necessarily regard that adjustment as precedent for
future cases, the circumstances in this case are very similar to the Idaho Power case.
lack the precise data to calculate a similar remedy of the mismatch in this case, but I note
that in the recent Idaho Power decision the Commission adjusted total revenues on the
order of 5 percent of the rate base additions.
Cost of Service Issues
HAVE YOU REVIEWED A VISTA'S COST OF SERVICE STUDY AND THE
RESUL TING RATE DESIGN?
Yes. The study sponsored by Ms. Tara Knox generally follows the methods approved in
the past, with a major exception described below. I recommend two improvements to
allocators contained in the Company s study.
A vista s Proposed "Four Factor" Allocator for Common Costs
DOES WITNESS TARA KNOX PROPOSE A CHANGE FROM THE PREVIOUS
APPROVED COST OF SERVICE METHODOLOGY USED IN CASE NO. WWP-
98-11 ?
DIRECT TESTIMONY OF DENNIS E. PESEAU - 33
IPUC Case Nos. AVU-04-1 and AVU-04-
Yes. As noted on Pages 6-7 of her direct testimony, the Company proposes to allocate
common costs" on the basis of four factors: direct O&M expenses, direct labor, net
direct plant, and number of customers. Previously, A vista had allocated these common
costs to customer groups with a 60% customer/40% energy allocation factor.
WHAT ARE "COMMON COSTS?"
Common costs are typically defined as those costs necessary for the utility to function
but which are left over after most directly assignable costs have been identified and
functionalized" to production, transmission, distribution or customer accounts. These
remaining common costs include general and common plant investment costs and
administrative and general expenses. Office buildings, furniture, transportation
equipment, certain inventories, computer costs and a portion of management salaries
typically comprise common costs.
ARE THE SPECIFIC FOUR FACTORS USED BY MS. KNOX TO ALLOCATE
COMMON COSTS P ARTIALL Y VALID?
Yes and no. Yes, the four factors, if correctly defined, are legitimate bases upon which to
allocate common costs. However, the method Ms. Knox uses to calculate the actual
weights of the four-factor allocations has a serious flaw, one that renders her calculations
highly volatile and incorrect.
PLEASE EXPLAIN.
In order to better explain this issue, I list the proposed four factors chosen for the
common cost allocations:
Direct O&M Expenses
Direct Labor Expenses
Net Direct Plant Expenses
Number of Customers
DIRECT TESTIMONY OF DENNIS E. PESEAU - 34
IPUC Case Nos. AVU-04-1 and AVU-04-
The issue I raise involves only one of the four factors - Direct O&M Expenses. Simply
put, Ms. Knox fails to remove a portion of these direct O&M expenses, an adjustment
that is necessary for the proper allocation of common costs.
WHAT ARE DIRECT O&M EXPENSES?
Direct O&M expenses in Avista s cost of service study are listed as FERC Accounts 500-
916 on pages 4-10 in Ms. Knox s Exhibit 16, Schedule 2. For reference, the sum of the
expenses in these O&M accounts is $97 699 000 (Line 369, Page 10 of 59, Exhibit 16
Schedule 2).
By using the sum of all the dollars in all of the O&M accounts, and their
allocators (energy, demand, customer) as one of the four factors used, Avista and Ms.
Knox are suggesting that common costs are caused in a fashion similar to the cause of the
O&M costs. Properly defined, O&M expenses form a reasonable means with which to
allocate common costs, but A vista s O&M expense definition fails in this regard.
WHAT IS THE BASIS FOR YOUR STATEMENT THAT A VISTA HAS
IMPROPERL Y DEFINED ITS DIRECT O&M EXPENSES AS ONE OF THE FOUR-
ACTORS TO ALLOCATE COMMON COSTS?
Three distinct reasons support my conclusion that Avista s first factor, the Direct O&M
Expense, incorrectly allocates common costs:
Avista s O&M expense allocator is extremely volatile from year to year
and common costs are not volatile.
A vista s annual common costs from 1998-2003 are actually inversely
related to its definition of O&M expenses.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 35
IPUC Case Nos. AVU-04-1 and AVU-04-
A statistical regression analysis supports the conclusion that the common
cost allocator using Avista s Direct O&M Expenses is valid if, and only if
variable fuel and purchased power expenses are removed.
A vista s Volatile Direct Expense Definition
WHAT IS THE ISSUE WITH RESPECT TO THE VOLATILITY OF USING
A VISTA'S DEFINITION OF DIRECT O&M EXPENSE TO ALLOCATE COMMON
COSTS?
Simply put, A vista s definition of O&M expenses includes fuel and purchased power
costs as an element from which the relatively fixed common costs are allocated. I offer
clear evidence below that common costs simply do not vary in any relation to changes in
fuel and purchased power costs.
AP ART FROM ACCOUNTING AND STATISTICAL DATA, IS THERE A COMMON
SENSE EXPLANATION AS TO WHY COMMON COSTS SHOULD NOT BE
ALLOCATED ON THE BASIS OF FUEL AND PURCHASED POWER COSTS?
Yes. As we are all aware, fuel and purchased power prices have risen, fallen, and again
risen by as much as several hundred percent on a year-to-year basis. Ifwe assume, as
A vista has done, that common costs are caused by changes in fuel and purchased power
costs, then we will be changing the common cost allocator by as much as several hundred
percent year-by-year.
Another way of stating the misapplication is that A vista is implying that its
expenditures on office buildings, furniture, parts inventories, vehicles, computers, office
supplies, employee pension and benefits, rents and general plant maintenance can be
expected to vary directly with the recent huge swings, both up and down, in fuel and
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
purchased power prices. (See Exhibit 16, Schedule 2, Pages 10-11 for complete list of
common (A&G) cost items.
DOES THIS DISTORT THE COST OF SERVICE RESULTS?
The distortion is huge, because fuel and purchased expenses from year to year comprise
the overwhelming majority of Direct O&M expenses. For example, of the total test year
O&M expenses of$97.7 million (Exhibit 16, Schedule 2, Page 10, Line 369) $66.
million, or 68 percent of the total is fuel and purchased power expenses. The effect on
customers of allocating relatively fixed common costs on volatile fuel and purchased
power prices is to cause huge swings in the levels of common costs allocated to each
customer class. These swings have nothing to do with the common costs of serving these
customer classes.
IS THERE AN EASY, COST -BASED FIX TO A VISTA'S VOLATILE AND
INACCURATE COMMON COST ALLOCATOR?
Yes, apart from the inclusion of fuel and purchased power expenses, the remaining Direct
O&M Expense factor is fairly indicative of, and related to the need to incur, common
costs. The easy fix is to simply remove the fuel and purchased power expenses and use
the remaining non-fuel and purchased power O&M expenses as one of the four-factors
for common cost allocator proposed by A vista.
Avista s Historical Common Costs are Inversely Related to Fuel
and Purchased Power Expenses
OTHER THAN YOUR COMMON SENSE DISCUSSION, HAVE YOU ATTEMPTED
TO ESTABLISH EMPIRICALLY THAT A VISTA'S EXPENDITURES FOR FUEL
AND PURCHASED POWER DO NOT DIRECTLY RELATE TO, OR CAUSE
A VISTA'S COMMON COSTS?
DIRECT TESTIMONY OF DENNIS E. PESEAU - 37
IPUC Case Nos. AVU-04-1 and AVU-04-
Yes. My Exhibit No. 204 is a graph of the recent history of Avista s annual variations in
total fuel and purchased power expenses comparing them with Avista s actual A&G
(common) costs, 1998-2003.
WHAT DOES EXHIBIT NO. 204 SHOW?
Exhibit No. 204 confirms what we know to be true - that Avista s fuel and purchased
power costs have varied tremendously on a year-to-year basis since 1998.
The exhibit also confirms the point I was making above, that Avista s common
(A&G) costs have been virtually constant since 1998. Use of the fuel and purchased
power expense component within A vista s Direct O&M factor would therefore generate
widely fluctuating allocations of common costs to different customer classes, distorting
the intent of a common cost allocator.
Statistical Relationship Between O&M and Common Costs
WHAT STATISTICAL VERIFICATION DO YOU HAVE THAT INDICATES THAT
A VISTA'S INCLUSION OF FUEL AND PURCHASED POWER EXPENSES IN ITS
COMMON COST ALLOCATOR IS INCORRECT?
The use of formal statistical analysis to prove that volatile, variable costs for fuel and
purchased power are not correlated with fixed common costs may be overkill, but I
nevertheless offer a statistical regression analysis supporting my arguments. The
statistical tests or "hypotheses" I review also indicate that fuel and purchased power costs
should be excluded from the allocator used to allocate common costs.
PLEASE EXPLAIN.
The regression analysis I performed simply answers the question of whether A vista
incurrence of common costs is fundamentally related to a definition of O&M expenses
DIRECT TESTIMONY OF DENNIS E. PESEAU - 38
IPUC Case Nos. A VU-04-1 and A VU-04-
that includes or does not include fuel and purchased power expenses. As our goal in the
cost of service study is to identify the causative factors of common costs, we search
statistically for the accounts making up O&M expenses that do, and those that do not
cause A vista to incur common costs. Then, in the allocation of common costs to
customer classes, we use only those O&M accounts that do relate to, or "cause" common
costs.
WHAT DOES YOUR STATISTICAL REGRESSION ANALYSIS SHOW?
The analysis shows that common costs are much more related to, or "correlated with
O&M expenses that have had fuel and purchased power expenses removed. The
regression analysis was conducted for two different equations:
Common Costs related to (O&M minus F&PP expenses); and
Common Costs related to (O&M with F&PP expenses)
where F&PP refers to fuel and purchased power.
Exhibit No. 205 summarizes the results of regressions for these two equations.
For completeness, common cost data were developed two ways: first measured as A&G
costs; second, as dollar levels of Avista s general plant accounts.
HOW WERE THE DATA DERIVED?
All data were taken from the 2003 FERC Form Is, for Avista and the five other western
electric utilities listed in Exhibit No. 205. The other five utilities provide a
representational cross section of similarly situated electric utilities.
PLEASE SUMMARIZE THE QUANTITATIVE FINDINGS.
Regardless of whether A&G expenses or general plant is used as the measure of common
costs, the regression results strongly indicate that O&M expenses less fuel and purchased
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
power expenses is a superior allocator, compared with Avista s proposed change of
including fuel and purchased power expenses. This analysis supports the common sense
reasoning and graphic evidence presented earlier, and it demonstrates that Avista
proposed change in these proceedings to include fuel and purchased power expenses to
allocate common costs should be rej ected.
HOW SHOULD COMMON COSTS BE ALLOCATED IN THESE PROCEEDINGS?
I believe that the Commission is left with two reasonable alternatives. First, the
Commission could adopt in principle Avista s four-factor common cost allocator concept
but simply order the Company to remove fuel and purchased power expenses from the
one factor, Direct O&M Expense. In this way, each of the factors in the four-factor
method would closely track common costs. I have participated in cost of service studies
in the past where FERC has similarly removed fuel and purchased power expenses from
the Direct O&M Expense accounts.
Alternatively, the Commission could order Avista to continue to use the
previously approved common cost allocator, where costs were allocated 40% on energy
and 60% on customer counts. The allocations resulting from the two alternatives are
similar in this case. My Exhibit No. 205 reflects the cost of service results from the four-
factor "Direct O&M less F &PP expenses" method.
My recommendation to the Commission is to use the four-factor Direct O&M less
F &PP expenses method.
A vista s Transmission Cost Allocator
DOES A VISTA'S COST OF SERVICE STUDY CORRECTLY ALLOCATE ITS
TRANSMISSION COSTS?
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
Transmission costs are incurred to meet peak demands, and are therefore appropriately
allocated to customer classes on the basis of demand (capacity) allocators. Avista
proposed cost-of-service study allocates a significant amount of transmission costs, not
on demand, but on an energy basis. This is no longer defensible.
DID A VISTA'S COST OF SERVICE STUDY IN WWP-98-AA ALLOCATE
TRANSMISSION COSTS SIMILARLY ON A DEMAND AND ENERGY BASIS?
Yes. Unlike the previous issue on the four-factor method, the transmission allocation
issue I raise here clearly would require the Commission to modify its position in the
previous rate case, and adopt the same methodology it recently approved in the Idaho
Power rate case. But I believe the evidence supporting this change is compelling.
PLEASE EXPLAIN.
My proposal to allocate transmission costs strictly on a demand basis is based on three
distinct propositions:
Avista s and virtually all other transmission systems are planned, sized
and built to meet maximum instantaneous, or peak demands.
A vista s proposed demand/energy transmission allocator is inconsistent
with, and contradictory to, the same transmission system rates it has had
approved, and indeed charges, to wholesale customers through its Open
Access Transmission Tariff ("OA TT"
The Commission has just weeks ago approved the demand allocator for
transmission costs that I propose here in the recently completed Idaho
Power general rate case.
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
WHAT IS THE BASIS FOR YOUR CONCLUSION THAT A VISTA'
TRANSMISSION SYSTEM IS CONSTRUCTED TO MEET ITS PEAK DEMAND
REQ UIREMENTS?
Our firm has examined system planning methods and models for many years. For
generation systems, a hydro-electric dam being a good example, construction costs can be
incurred to meet both demand and energy considerations. In the Pacific Northwest, for
example, we know that hydro generation costs are incurred or "caused" not only by peak
demand requirements, but also by the need to store energy. Generation costs are
routinely allocated to both demand and energy.
Transmission systems, while they obviously transmit energy, are planned for, and
the cost is caused by, the need to meet peak demands. Once the costs are incurred and
the facilities constructed, no additional costs are incurred to transmit energy. Thus, the
principle of cost-causation leads us to allocate transmission on the basis of demand
(capacity) usage only.
HOW IS A VISTA'S PROPOSED DEMAND/ENERGY TRANSMISSION
ALLOCATOR INCONSISTENT WITH THE TRANSMISSION COST ALLOCATION
AND RESULTING RATES IT HAS IN PLACE FOR WHOLESALE TRANSMISSION
USERS?
The open access policies implemented by FERC some years ago, as we know, require
A vista and other utilities to file and maintain OA TTs, the rates of which must be based
on cost of service. I have reviewed the current A vista OA TT and determined that the
Company allocates its transmission system costs (the same system contained in its
present transmission cost of service) not on the basis of the demand/energy allocator
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
proposed in this general retail rate case, but rather on the same demand basis that I am
proposing here. There is no reasonable justification to have two different sets of
transmission costs and rates for the same identical system.
HOW DO YOU KNOW THAT THE APPROVED OATT RATE IS BASED ON A
DEMAND-ONL Y ALLOCATOR?
In my Exhibit No. 207 I attach a copy of the relevant pages of Avista s present OATT.
The rates posted there are derived strictly on a "per kW" or demand basis. This indicates
that the OA TT rates and the transmission costs contained therein are based solely on a
demand allocator.
DO PROBLEMS ARISE FROM ALLOCATING THE SAME TRANSMISSION
COSTS OF SERVICE ON THE BASIS OF TWO DIFFERENT ALLOCA TORS, AS
VISTA IS PROPOSING?
Yes, obviously so. First, the demand method is correct and the demand/energy is not.
Therefore, one set of rates is correct and the latter is not. There is no sound reason why
identical retail or wholesale transmission customers should have their respective cost
allocations and therefore their rates differ for the same usage. This is disparity is not only
illogical; it is also potentially discriminatory.
WHAT TRANSMISSION COST ALLOCATION METHOD DID THIS COMMISSION
ADOPT IN THE RECENT IDAHO POWER GENERAL RATE CASE NO. IPC-03-13?
The Commission based its rate design on Idaho Power s basic cost of service study,
which allocated the Company s transmission costs on the basis of demand only. Idaho
Power s approved OATT rates are also based on demand-only transmission cost
allocators.
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
HAVE YOU PREPARED A COST OF SERVICE STUDY THAT INCORPORATES
THE CHANGES YOU RECOMMEND?
Yes. Exhibit 206 is a summary of the results of my cost of service study incorporating the
proper 4-factor and transmission capacity allocator. While the changes to the allocations
to the various customer classes are not dramatic, they are significant and necessary to
properly capture cost of service.
WHAT DOES YOUR COST OF SERVICE STUDY SHOW WITH RESPECT TO THE
PRESENT CONTRIBUTIONS THAT DIFFERENT CUSTOMER CLASSES ARE
MAKING TOWARD RESPECTIVE COSTS OF SERVICE?
The summary results indicate, consistent with the conclusions of A vista s cost of service
study, that residential customers, Schedule 1 , and large general service customers
Schedule 25, are receiving substantial subsidies from all remaining customer classes
including Potlatch. Page 1 of Exhibit 206 shows that the residential and general service
customer classes' rates generate rates of return that are significantly below the system
average rate of return, thus indicating that other classes ' rates are set too high in order to
make up the shortfall.
HOW SHOULD THE COMMISSION DEAL WITH THE ELIMINATION OF THESE
SUBSIDIES?
In the recent Idaho Power general rate case I testified that a huge subsidy, in that case to
the irrigation pumping class, needed to be systematically and unequivocally reduced to
zero, necessitating a large increase to the irrigators. The same principles apply here
although the levels of subsidies to the residential and general service customers are not so
large as in the Idaho Power case. In principle, I believe these subsidies should be
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. A VU-O4-1 and A VU-04-
eliminated immediately. However, I am also aware the Commission has expressed
concerns about the "rate shock" that could result from very steep increases for a
particular customer class.
Accordingly, I propose in these proceedings that, if the overall approved increase
is ten percent or less, all customer classes should be moved to full cost of service. If the
increase is greater than ten percent, the Commission order should order residential and
large general service rates moved at least halfway toward rate of return parity, with two
annual automatic adjustments thereafter to close the remaining cost of service gap.
Under the latter alternative, the other customer classes (Schedules 11-, Schedules 21-
, and Potlatch) would continue to pay a subsidy in the near term, but would receive
assurances that the remaining subsidy would be eliminated over the next two years. This
, I believe, more than fair to the subsidized customer classes.
Rate Design Issues
DO YOU HAVE ANY COMMENTS ON A VISTA'S RATE DESIGN PROPOSALS?
Yes. My first observation is that Avista s proposal to include Potlatch's Lewiston
Facility ("Facility ) in Tariff Schedule 25 should be rejected. Because of the huge
disparity in size between the Facility and the other Schedule 25 customers, it makes no
sense to include the Facility in that schedule. For customers the size of the Facility, the
Commission has always used separate tariffs for each special contract customer, and it
should do so in this case as well. The Facility is approximately three times the size of all
the entire Schedule 25 class.
IS THE FACILITY IN FACT A SPECIAL CONTRACT CUSTOMER?
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. AVU-04-1 and AVU-04-
Yes. The A vista and Potlatch power supply agreement ("Agreement") is a unique
contract that governs Avista s service to only one customer- the Facility. In that
Agreement, the parties agreed to the temporary use of Schedule 25 rates for service to the
Facility, pending the next rate case. But Potlatch did not agree to become a Schedule 25
customer The Facility has always been a "special contract customer" in the past, and the
Agreement clearly contemplates that this status will continue in the future.
IS IT DIFFICULT TO SEPARATE THE FACILITY'S COST OF SERVICE FROM
SCHEDULE 25?
No. The A vista cost of service study, and my own, already compute all cost of service
elements for the Facility on a stand-alone basis, in recognition of the fact that the Facility
is indeed a customer class unto itself. Given this, the Commission should require A vista
to preserve these cost elements treating the Facility as the customer class that it is.
makes no sense to subsequently meld the Facility with the much smaller Schedule 25
class. In order to set rates for the Facility within the Schedule 25 class, Avista in this
case had to resort to major rate design changes in order to properly assure that Potlatch
would not be overcharged.
Creating a stand-alone rate schedule for the Facility will not affect the Facility
cost of service or rates. It is simply a preventive measure. The concern is that in the
future this distinction could be blurred in a subsequent study in a manner that causes the
Facility to pay costs for which it should not be accountable. The distinction between the
Facility and the Schedule 25 customers should be clarified by placing the Facility in a
separate rate schedule.
DOES THIS COMPLETE YOUR TESTIMONY?
DIRECT TESTIMONY OF DENNIS E. PESEAU - 46
IPUC Case Nos. A VU-04-1 and A VU-04-
Yes, it does.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 47
IPUC Case Nos. A VU-04-1 and A VU-04-
Appendix A-Update to Dr. A vera s Analysis
WHAT IS THE CORRECT RETURN ON EQUITY RANGE USING DR. AVERA'
METHODS FOR ESTIMATING EQUITY RETURNS?
I conclude that consistent application of the discounted cash flow (DCF) and risk
premium methods used by Dr. Avera reduces his recommendations as follows:
ROE Method Avera Estimate-Peseau U date
DCF 10.4%
Risk Premium I 11.4 10.
Risk Premium II 10.9.2% to 10.
CAPM 11.9 10.
- n/ includes flotation costs of 20 basis points.
Updates that are consistent with the methods Dr. Avera utilizes do not support his range
of 10.4% to 11.9% and certainly do not support a recommended ROE of 11.5%. See
Exhibit No. 211.
WHAT GENERAL COMMENTS DO YOU HAVE REGARDING THE TESTIMONY
AND ANALYSES OFFERED BY DR. AVERA?
Dr. Avera offers 70 pages of testimony covering a number of topics. Twenty-four of
these pages cover discussion of flotation costs and the quantitative equity return methods
and estimates commonly considered by this Commission. The rest of the testimony is
concerned with general and fundamental economic and financial topics that are normally
and efficiently taken into account by investors when bidding on and purchasing common
stock and other assets. Financial institutions and investors know the financial and
operational characteristics of A vista every bit as well as Dr. A vera and use this
information to make formal investment decisions. A well-known financial principle is
that investors are not normally, nor do they expect to be, compensated for nonmarket or
DIRECT TESTIMONY OF DENNIS E. PESEAU - 48
IPUC Case Nos. A VU-04-1 and A VU-04-
company-specific risks that are not systematic. These risks are diversifiable and do not
and should not form the basis of rate of return "adders." The methods of determining
cost of equity used by Dr. Avera and others in this case measure returns that are
commensurate with similar risk-adjusted investments and should not be adjusted for
exogenous risks.
PLEASE SUMMARIZE DR. AVERA'S ESTIMATES.
Dr. Avera presents four quantitative analyses of the cost of equity for a "benchmark"
group of western electric utilities from which he derives a 10.2% to 11.7% equity cost
range. He presents a discounted cash flow ("DCF") analysis for a benchmark group of
electric utilities in the western U. S., two risk premium approaches, and an estimate based
on the capital asset pricing model ("CAPM"
).
From his DCF analysis, he estimates that a
benchmark sample of western electric utilities requires a return on equity of 10.2% (page
45). Based on two risk premium models, he concludes that the cost of equity for the
respective reference samples of electric utilities is 11.2% (page 49) and 10.6% (page 50).
And, from his CAPM approach, he derives a cost of equity estimate for the western
electric utilities of 11.7% (page 51). Based on that information, and an adder of 20 basis
points for flotation costs and additional premiums he argues are required for risk specific
to A vista, he endorses an ROE of 11.5%.
HOW DOES HE REACH THE CONCLUSION THAT A VISTA SHOULD BE
AUTHORIZED AN EQUITY RETURN IN EXCESS OF 11.5%?
Dr. A vera presents lengthy discussions of company-specific risks that he contends are
faced by A vista and should be recognized in setting the authorized return. That analysis
of unique risks is the basis for his contention that the Company requires an equity return
DIRECT TESTIMONY OF DENNIS E. PESEAU - 49
IPUC Case Nos. AVU-04-1 and AVU-04-
near the top of his estimate of the equity cost range for other western electric utilities.
But as I just explained, these company specific risks are incorporated into his results, and
a subjective adder for such risks is unwarranted.
Update to Dr. Avera s DCF Approaches
DO YOU HAVE ANY COMMENTS ABOUT HIS DCF ANALYSIS?
Yes. Recall that the DCF method under standard financial assumptions reduces to the
equation:
ROE = D1/Po + g
where ROE required equity return
first period dividend rate
today s stock price
growth rate
Dr. Avera s estimate of a 10.2% return results from his estimate of the DCF components:
10.2% = 4.2% (yield) + 6.0% (growth)
I update the 6.0% growth rate and his dividend yield. The growth rate g is growth that is
expected in the future by investors. It is by nature forward looking. But I note that on
Dr. Avera s Schedule WEA-, he used not only the typical benchmark for expected
growth, as reported by the investor institutions IBES, Value Line, First Call and Multex
Investor, but also historical rates of earnings growth for both five and ten year past
periods:
DIRECT TESTIMONY OF DENNIS E. PESEAU - 50
IPUC Case Nos. A VU-04-1 and A VU-04-
Dr. Avera s Ex ected Growth Rates
Value First Past Past
IBES Line Call Multex 10 Yr.5 Yr.
Average Expected
Growth Rate 5.1 2.4 5.4 7.3
While the simple average of these growth rates is 5., Dr. Avera inexplicably uses a
0% figure to develop his 10.2% return.
IN YOUR OPINION, IS DR. AVERA'S USE OF THE HISTORICAL GROWTH
RATES IN HIS AVERAGE AN APPROPRIATE BASIS FOR ESTIMATING THE
DCF REQUIRED FUTURE EXPECTED GROWTH RATE?
No. To the extent that past growth might be of any importance to investors, the analysts
forecasts Dr. Avera reports for IBES , Value Line, First Call and Multex have already
taken that information into account. David A. Gordon, Myron J. Gordon and Lawrence I.
Gould
, "
Choice Among Methods of Estimating Share Yield Journal of Portfolio
Management pp. 50-55 (Spring 1989), did a study that found analysts' forecasts of
growth provide a better explanation of stock prices than three backward-looking
measures of growth. They explain that their findings make sense because analysts would
take into account past growth as well as any new information when they form their
forecasts. Roger Morin reports the results of other empirical studies and concludes
analysts' forecasts "are more accurate than forecasts based on historical growth.
Regulatory Finance: Utilities Cost of Capital, page 154. My restatement of Dr. Avera
DCF analysis recognizes four of the growth forecasts Dr. Avera relied upon, but gives no
weight to the measures of past growth Dr. Avera reported.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 51
IPUC Case Nos. AVU-04-1 and AVU-04-
HOW HAVE YOU MODIFIED DR. AVERA'S DCF EXPECTED GROWTH RATE
ARIABLE TO REMOVE THE EFFECTS OF HISTORICAL GROWTH?
My Exhibit No. 208 shows those results. To determine an updated and consistent
estimate for the DCF expected growth rate for each of the utilities in Dr. Avera s sample
I updated his reported estimates of investor institution projections in Schedule WEA-2 as
well as his estimate of sustainable growth in his Schedule WEA-3. Exhibit No. 208
shows an average of four growth forecasts; the current estimates reported by IBES, First
Call and Reuters (formerly Multex) and the higher of the two forecasts made with Value
Line data. Exhibit No. 208 shows that the correct average for the projected or expected
growth rate is 5.1 %, close to the bottom of the 5% to 7% range adopted by Dr. Avera.
DID YOU UPDATE DR. AVERA'S DIVIDEND YIELDS?
Yes. I used data published by Value Line, dated June 4, 2004, and the method Dr. Avera
used to compute dividend yields to make that update. These updated dividend yields are
also reported in Exhibit No. 208.
BASED ON YOUR UPDATES AND UTILIZATION OF ONLY THE FORWARD-
LOOKING GROWTH RATES REPORTED BY DR. AVERA, WHAT IS YOUR
RESTATEMENT OF DR. AVERA'S DCF RESULTS?
Based on his sample and the restatements discussed above, the indicated average cost of
equity for the western electric utilities is 9.3% (4.% dividend yield and 5.% growth
after rounding), 90 basis points less than the 10.2% estimated by Dr. Avera.
DO YOU HAVE OTHER CONCERNS WITH DR. AVERA'S DCF ANALYSIS?
Yes. The DCF method he proposes is incorrect. At page 32, Dr. Avera presents the
general form of the DCF model. It clearly shows that expected dividends per share
DIRECT TESTIMONY OF DENNIS E. PESEAU - 52
IPUC Case Nos. AVU-04-1 and AVU-04-
(DPS) are the cash flows that are of interest to investors. He adopts Value Line
forecasts of dividends for the next year but ignores Value Line ~ forecasts of dividends for
other future years. His DCF approach is incorrect because it does not incorporate all of
the information on dividend growth that investors consider when they price the shares of
common stock in his sample. Had Dr. Avera made his DCF estimates with a multi-stage
DCF model that recognized that dividend growth is expected to be less than half as rapid
as forecasted earnings and sustainable growth for the period 2004 to 2008, the DCF
equity cost estimate would be less than 9.3%. But because I limit my testimony to a
restatement of the methods Dr. Avera has relied upon, I have not presented such an
analysis.
Update to Dr. Avera s Risk Premium Approaches
PLEASE DESCRIBE THE RISK PREMIUM APPROACH TO ESTIMATING A
UTILITY'S REQUIRED RETURN ON EQUITY.
Whereas the DCF method adds estimates of dividend yield to expected growth rate to get
equity cost estimates, risk premium methods recognize that over time common stock is
riskier than most debt securities (bonds) and therefore requires a premium, or adder, over
and above the return on bonds. This adder is often termed a risk premium. As yields on
bonds are generally directly observable and measurable, equity cost estimates may be
derived if reliable risk premiums can be determined.
HOW DOES DR. AVERA UTILIZE THE RISK PREMIUM METHOD?
Dr. Avera uses a risk premium method based on authorized equity returns, another based
on actual or realized returns and, finally, the more academically rigorous risk premium
method, the Capital Asset Pricing Model (CAPM).
DIRECT TESTIMONY OF DENNIS E. PESEAU - 53
IPUC Case Nos. AVU-04-1 and AVU-04-
WHAT EQUITY RETURN DOES DR. AVERA ESTIMATE USING HIS
AUTHORIZED RETURN RISK PREMIUM METHOD?
11.2%. He derives this by adding a December 2003 bond yield of 6.61 % to a risk
premium estimate of 4.58% that is derived in his Schedule WEA-4. Schedule WEA-
uses regression analysis to attempt to determine the historical relationship between
allowed equity returns and bond yields, and the difference between the two, to establish
the risk premium. The theory is that if the regression analysis can determine the
relationship between the bond yield and the appropriate risk premium, then one can
observe today s bond yield, add to it the estimate of risk premium appropriate for the
bond yield and add the two to get an equity return estimate. From Schedule WEA-, Dr.
Avera estimates the relationship as:
(ROE - Bond Yield)
= .
073 + (-.435 x Bond Yield)
While I have no quarrel with the basic methodology, Dr. Avera uses interest rates or bond
yields that are internally inconsistent in his method.
PLEASE EXPLAIN.
Dr. Avera uses a low yield bond to compute his historical risk premium. Use of this low
bond yield when subtracted from allowed equity returns, produces an exaggerated or
higher risk premium than if a consistent bond rate is used. The bond yield used by Dr.
Avera, shown on Schedule WEA-4 is an average of AAA, AA, A and BBB rated bonds.
Since the highly rated bonds AAA, AA and A will have the lowest interest rates, the
composite rate Dr. Avera uses is low. Subtracting a low interest rate from an authorized
return yields an artificially high risk premium. Then, on Page 49, Line 10, he adds this
high risk premium to the highest bond yield, that of a triple-B bond. This mixing of
DIRECT TESTIMONY OF DENNIS E. PESEAU - 54
IPUC Case Nos. A VU-04-1 and A VU-04-
different bonds for the regression analysis and for computing the equity return biases
upward Dr. Avera s estimate of an equity return.
HAVE YOU ATTEMPTED TO REMOVE DR. AVERA'S INCONSISTENCY?
Yes. An appropriate calculation would use the same measure of bond rating in the
regression analysis as in the recommended equity return. In making my restatement, I
have used A-rated utility bonds to compute the risk premiums, to run the regressions and
to estimate the equity cost. I ~hose the A-rated utility bond rates because Dr. Avera relies
on A-rated bonds in Schedule WEA-5. Also, current quotations for A-rated utility bond
rates are widely available and published by Value Line every week. I also used triple-
rates, as a second approach in another regression as well, because that is what Dr. Avera
uses on his Page 49.
The results of the revised analysis are shown in my Exhibit No. 209, pages 1 and
2. Combining the revised regression result with a June 4 2004 Value Line quotation of
08% for A-rated utility bond rates gives an indicated cost of equity for the benchmark
electric utilities of 10.40 basis points lower than Dr. Avera s estimate of 11.2%.
Using the triple-B regressions with the current triple-B rate of 6.56% reported June 4
2004 gives a cost of equity estimate of 10.9%.
DO YOU HAVE ANY COMMENTS ABOUT DR. AVERA'S RISK PREMIUM
APPROACH BASED ON THE REALIZED-RA TE-OF -RETURN APPROACH THAT
HE PRESENTED IN SCHEDULE WEA-
Yes. First, as he did with his other risk premium approach, Dr. Avera used one type of
bond to determine the average risk premium and then incorrectly added that risk premium
to a triple-B public utility bond rate. In this analysis the risk premium was established as
DIRECT TESTIMONY OF DENNIS E. PESEAU - 55
IPUC Case Nos. AVU-04-1 and AVU-04-
the average difference between annual returns on stocks and A-rated bonds and thus the
risk premium will be larger than if the premium were established for triple-B bonds. To
make Dr. Avera s approach internally consistent, I added the current A-rated bond to the
premium for A-rated bonds. This change alone reduces Dr. Avera s equity cost estimate
to 10.
%.
See Exhibit No. 210.
My other observation is that Dr. Avera s approach assumes that investors
typically have holding periods of only one year, when investors probably expect to hold
shares of utility stocks for longer periods. If investors have very long holding periods, a
risk premium based on differences in geometric average returns would be the appropriate
risk premium. If, for example, investors have 57-year holding periods, the correct
estimate of the risk premium would be 3.11 % instead of 4.01 %. See Exhibit No. 210. I
expect that investors typically have holding periods longer than one-year but much
shorter than 57 years. In such a case this approach would indicate the cost of equity
would be between 9.2% and 10.% but closer to 10.
DO YOU HAVE ANY COMMENTS ABOUT DR. AVERA'S CAPITAL ASSET
PRICING MODEL EQUITY COST ESTIMATE?
Yes. Although the CAPM's derivation is steeped in a good deal of financial theory and
mathematical determination, the final specification, like the DCF method, is fairly
straightforward:
Equity Cost = Risk Free Rate + Beta x Market Risk Premium
There are a number of different ways the CAPM can be implemented and a number of
ways that estimates of the risk free rate and market risk premium can be derived. I limit
DIRECT TESTIMONY OF DENNIS E. PESEAU - 56
IPUC Case Nos. AVU-04-1 and AVU-04-
my comments to an update of Dr. Avera s risk free rate and his estimate of the market
risk premium (MRP). I will not contest his measure of market risk
, "
beta.
WHAT IS THE RISK-FREE RATE USED BY DR. AVERA?
Dr. Avera uses as a measure of the risk-free rate the average yield on long-term
government bonds. He indicates that this measure of the risk-free rate as of December
2003 was 5.2%.
WHAT IS THE RECENT YIELD ON LONG-TERM GOVERNMENT BONDS?
The yield reported by Value Line at June 4 2004 is 5.32%. I use that value in my update
of Dr. Avera s CAPM estimate.
HOW DOES DR. AVERA ESTIMATE THE MARKET RISK PREMIUM ("MRP"
While I do not agree with his method of estimating the MRP, I use his method here with
a simple update.
Dr. Avera derives a forecast of the total average market return for the stock
market of 13., then, to estimate the market premium he subtracts his risk free rate of
5.2%, which results in an 8.5% MRP.
WHAT UPDATE HAVE YOU MADE TO DR. AVERA'S MRP?
Whereas the long-term government bond rate is directly observable and is set in
competitive markets, the other component of the risk premium approach used by Dr.
Avera, the projected market return, is not directly observable or measurable. The
projected market return is simply the opinion about the future made by different investor
institutions and can change frequently. Use of a projected market return of 13., as of a
single point in time, therefore makes the prediction of total market return highly variable
as I now show. For reference, the long-term average market risk premium during the
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. A VU-04-1 and A VU-04-
period 1926 to 2003 is 7., not the 8.5% used by Dr. Avera. Investors that use CAPM
would undoubtedly give weight to that long-term average market risk premium.
Dr. Avera s total market return estimate was made prior to recent stock market
activity that has occurred since December 2003. Investors now understand that a short-
term gain as large as 13.7% is no longer realistic. For example, the Value Line forward-
looking total market return for the 1700 stocks it follows, as of June 4, 2004, was
12.55%, not the 13.7% used by Dr. Avera. This huge potential for variation in these
current" MRP estimates makes rate of return setting for regulatory purposes difficult.
Nevertheless, using the updated market return forecast of 12.55%, the implied MRP is
23% (12.55% - 5.32%), not the 8.5% used by Dr. Avera. At this time, the indicated
current" market risk premium and the long-term average market risk premium are both
2%. If investors consider either indicator of the market risk premium, an update of Dr.
Avera s CAPM equity cost estimate is 10.9% as shown below:
Equity cost RF + beta x MRP
Equity cost = 5.32% + .77 x 7.2% = 10.
PLEASE SUMMARIZE YOUR UPDATES AND RESTATEMENTS OF DR.
AVERA'S QUANTITATIVE ESTIMATES OF THE COST OF EQUITY FOR
BENCHMARK ELECTRIC UTILITIES.
I conclude my straightforward updates of Dr. Avera s estimates of the cost of equity do
not support a recommended ROE range of 10.4% to 11.9% and certainly do not support
an equity return for A vista of 11.5%. My summary Schedule DEP-4 shows that a simple
average of the updated equity cost estimates is 140 basis points below the 11.5% ROE
that Dr. Avera recommends for A vista.
DIRECT TESTIMONY OF DENNIS E. PESEAU - 58
IPUC Case Nos. A VU-04-1 and A VU-04-
DO THE DIRECTIONS IN TRENDS OF FINANCIAL MARKETS SUPPORT YOUR
RECOMMENDATIONS?
Yes. My Exhibit No. 212 shows monthly interest rate data for 10-year Treasury bonds
and for Baa corporate bonds for the period October 2001 through April 2004, as reported
by the Federal Reserve. Generally, rates for government bonds and Baa corporate bonds
have decreased by 145 basis points since October 2001. I conclude that, given the drop
in capital costs, A vista s cost of equity is well below its 1998 cost.
DIRECT TESTIMONY OF DENNIS E. PESEAU -
IPUC Case Nos. A VU-04-1 and A VU-04-